Foreign language learning apparatus and method for correcting pronunciation through sentence input

Abstract

A foreign language learning apparatus and method are provided. The foreign language learning apparatus includes a sentence input unit receiving a first sentence from a user; a linked letter detection unit detecting at least one letter corresponding to at least one linking rule; a linked letter removal unit removing the letter and generating a second sentence by inserting a linking code; a partial waveform generation unit generating one or more partial waveforms using the Text To Speech (TTS) engine; an input waveform generation unit converting a voice corresponding to the first sentence into an input waveform; and a matching degree calculation unit calculating a matching degree and a partial matching degree. This foreign language learning apparatus enables a user to effectively learn pronunciation of a foreign language.

Description

CROSS REFERENCE[0001]This application claims foreign priority under Paris Convention and 35 U.S.C. §119 to Korean Patent Application No. 10-2013-0137779, filed Nov. 13, 2013 with the Korean Intellectual Property Office.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]The present invention relates generally to a foreign language learning apparatus and method for correcting pronunciation through sentence input and, more particularly, to a foreign language learning apparatus and method for correcting pronunciation through sentence input, in which a waveform is disposed in accordance with a sentence input by a user and a matching percentage is calculated by matching a voice input by a user with the previously stored waveform or a waveform generated via a Text To Speech (TTS) engine.[0004]2. Description of the Related Art[0005]The present invention relates to a foreign language learning apparatus.[0006]Recently, in line with the trend of the specialization and globalizatio...

AI Technical Summary

Benefits of technology

The present invention is a foreign language learning apparatus and method that uses sentence input to correct pronunciation. It generates a curriculum for words that have a low match rate with a user's voice input and uses that curriculum to effectively correct pronunciation.

Problems solved by technology

In such a case, problems arise in that high costs are required for English learning and the participation of people who lead a busy life, such as office workers, in the learning of English is very limited because education is performed according to specially arranged schedules.

However, disadvantages arise in that it is difficult to expect correct evaluation for the characteristic data extracted via the AM or from the voice because errors attributable to the personal pronunciation propensities of all native speakers are disregarded and the stress or isochronism of a syllable or sentence structure and a sentence cannot be evaluated or the accuracy of the evaluation cannot be expected even when the stress or isochronisms are evaluated.

Furthermore, the above-described pronunciation comparison and analysis programs are problematic in that they may not be used to evaluate individual characteristics, such the accent, stress, or pronunciation speed of a specific sentence or word, because the same or random weight is equally assigned to the segmental and non-segmental characteristics of all types of pronunciations.

Furthermore, a learner may be roughly aware of the degree of accuracy of his or her pronunciation indicated by the scores, but may not learn based on accurate comparison and analysis because there is no means for making a comparison on pronunciation for each field, such as each of the pronunciation length, accent, stress and pronunciation speed of a word or a sentence.

As a result, a problem arises in that pronunciation correction is limited.

Furthermore, conventional language learning machines on which voice recognition programs or engines have been installed include only uniform evaluation criteria.

Accordingly, conventional language learning machines are unable to enable learning suitable for the level or personality of a learner because the language ability or pronunciation characteristic of a learner are not appropriately reflected in the language learning machines.

Furthermore, it is difficult to apply conventional voice recognizers to the AM of a foreign language native speaker, and a recognition ratio for the accuracy of pronunciation falls short of learners' expectations due to different pronunciation and language habits.

Furthermore, it is practically impossible to effectively distinguish minimal pairs, that is, an elementary learning target using conventional voice recognizers, it is impossible to effectively handle a noisy environment, and the detection performance of a keyword or core words and phrases is low or ineffective.

Furthermore, current language learning machines are problematic in that they cannot interpret and handle various language phenomena that may occur in actual conversation environments because the current language learning machines are focused on textbook learning methods.

Furthermore, conventional language learning machines are problematic in that they are expensive.

Furthermore, there is no provision for a system for accurately evaluating a foreign language pronounced by a learner in various ways, providing notification of the differences between a learner's pronunciation and a native speaker's pronunciation, and digitizing the results of the evaluation.

Images